"Model spatial heterogeneity in environmental and ecological processes"
企画者 中谷 友樹 / 東北大学大学院 環境科学研究科
Spatial statistics have been widely adopted in ecological analysis including fitting models of environmental and epidemiological processes to empirical datasets. This symposium focuses on techniques to infer spatial heterogeneity of processes as one of essential aspects of spatial statistical modelling with ecological applications. Geographical relationships between spatially observed variables often vary over space indicating spatial heterogeneity in processes to generate spatial phenomena. Geographically weighted regression and related modelling approaches are commonly used for systematically estimating detailed geographical variations in parameters to be estimated in various multivariate modelling. This symposium intends to highlight the concepts of such spatial statistical approaches and how ecological and environmental analysis can be extended by them as well as their challenges in the era of big spatial data science.
"Geographically weighted modelling: principles and applications in spatial epidemiology" 中谷 友樹 / 東北大学大学院 環境科学研究科
"Balancing spatial and non-spatial heterogeneity in large samples" 村上 大輔 / 統計数理研究所 データ科学研究系構造探索グループ
"Spatial heterogeneity of errors in land cover data" 堤田 成政 / 京都大学大学院 地球環境学研究科
"Spatial models for compositional data: Considering spatial auto-/cross-correlation and heterogeneity" 吉田 崇紘 / 国立環境研究所 地球環境研究センター